Audit of what already runs
We look at what is in place, how it is connected and what it costs. Model spend is priced separately from infrastructure.
03 / CONSULTING
AI, software and IT consulting
We work out why a deployed system is not delivering. Most often the model is fine and the problem sits in what gets measured, and where. The same discipline applies to software and IT decisions in general.
The typical picture: the system is live, the money is spent, and the effect is nowhere to be seen. The cause is rarely model quality. Usually something in the measurement is broken: an event fires on the wrong step, a metric counts the wrong thing, a report shows a number that affects nothing.
An audit starts by checking what is measured at all, and only then moves to the solution itself.
The same approach works beyond AI. Choosing a stack, a hosting model or a vendor, deciding whether to build or buy, reviewing an architecture before a large spend: these are software and IT decisions, and we take them the same way, numbers first.
Measurement check
We verify that events fire where they should and that the numbers in the report reflect reality.
Where AI is redundant
We mark the places where the model can be replaced by a rule or plain code: cheaper, faster and more predictable.
Software and IT decisions
Stack, hosting, build or buy, vendor review, architecture review before a large spend. Each decision comes with the cost of running it, on paper, before it is made.
Work plan
A prioritised list of changes: what pays off at once, what needs rework, what can be left alone.
What this is based on
A telling case from our own practice: an ad account showed zero conversions on a key goal for months. The problem was neither the ads nor a model. The event fired on a step that users never reached: people got their result and left, and the system counted no result at all. Moving the event to the right step restored the real picture, and only then could the campaigns be optimised.
What you get
- A report: what works, what does not, what costs money for nothing
- Corrected measurements you can trust
- A prioritised work plan with an estimated effect for each item
- A real figure for the cost of operation
When it is not needed
- If there is no solution yet, there is nothing to audit; that is a design task
- If you need a formal document for reporting rather than changes
If the task is not for us, we say so in the first conversation and, where we can, point you to someone who fits.
Questions we get asked
What do you need from us for an audit?
Access to what already runs, and one person who knows the process from the inside. Without the second, an audit turns into reading code with no idea why it was written.
Do you work with systems you did not build?
Yes, that is most of our audits. We read other people's code rather than rewrite it: the goal is to understand what is happening, not to replace everything with our own.
How much does it cost to operate an AI solution?
It depends on load and model choice, and the range is tens of times wide. Part of the audit is calculating that figure for your case and showing where it comes down without losing quality.
Do you advise on IT questions that have nothing to do with AI?
Yes. Stack and hosting choices, build versus buy, vendor and architecture reviews. The method is the same: what it costs to run, what breaks, how it rolls back.
Other services
01 / DESIGN AI Design We work out the economics before the first line of code. Half of the AI requests we see are solved simpler and cheaper, and that is best known at the start. Read more 02 / DEV AI Development We build systems that survive a provider outage, a load spike and a model change. Monitoring ships with the release, and every change has a rollback plan. Read more
Tell us about the task
We will look at where AI pays off and where it would cost more than it returns.